Abstract
It has been argued that self-generated explanation (self-explanation) is an important tool in the acquisition of knowledge. However, the rationalist and the empiricist approaches, with which self-referential activity has usually been discussed, makes the phenomenon of self-explanation problematic. In this article, I examine self-explanation as a private case of the classical learning paradox, which is associated with self-referential activity, and consider an activity oriented solution to this difficulty. From a practical perspective, I suggest a reframing of self-reference in Activity Theory in order to provide researchers with a method for analyzing and supporting bootstrapping phenomena, such as the self-explanation effect, in learning and problem solving.